Business

We choose to limit the scope to our recommender system to only businesses with tags related to food and beverages. There were originally 508 unique category tags listed within our business data. We manually filtered 112 targeted categories to subset our data.

We applied additional transformation to remove unnessacary data. There were 1,224 business in our data that were permanently closed. These companies accounted for 9.8% of all businesses, which were subsequently removed from our data. There were also 3 businesses in our dataset from outside of AZ that we also removed.

As a result of our transformations, our recommender data was shortened 4,828 unique businesses. This was further limited to 4,332 after randomly sampling our user-data. The output of which can be previewed below:

Review

We subset our review data from the subset of food and beverage businesses. This dropped our review data from 229,907 to 165,823 reviews. We later applied another filter to the data to only use reviews from 10,000 randomly sampled users. This further decreases reviews to 44,494 observations. Our review data can be previewed in two parts below:

User

Last, we applied a similar filter to users to subset our data based on only our selected businesses. This decreased our user data from 43,873 to 35,268 distinct user_id observations. Do to processing constraints in R, we choose to randomly sample 10,000 users from these unique profiles.

The dataframe preview below shows aggregate user data for all reviews an individual user provided for yelp within our data selection.

User-based CF

UB <- Recommender(getData(evalu, "train"), "UBCF", param = list(normalize = "Z-score", 
    method = "Cosine"))

p <- predict(UB, getData(evalu, "known"), type = "ratings")

p@data@x[p@data@x[] < 1] <- 1
p@data@x[p@data@x[] > 5] <- 5

calcPredictionAccuracy(p, getData(evalu, "unknown"))

FALSE     RMSE      MSE      MAE 
FALSE 1.384047 1.915586 1.018311

Item-based CF

IB <- Recommender(getData(evalu, "train"), "IBCF", param = list(normalize = "Z-score", 
    method = "Cosine"))

p1 <- predict(IB, getData(evalu, "known"), type = "ratings")

p1@data@x[p1@data@x[] < 1] <- 1
p1@data@x[p1@data@x[] > 5] <- 5

calcPredictionAccuracy(p1, getData(evalu, "unknown"))

FALSE      RMSE       MSE       MAE 
FALSE 1.3710486 1.8797743 0.9719316

Connecting

We initiated a local connection with Spark (V2.4.3). Our yelp data was inputted into a spark table and split for training and testing purposes.

# configure spark connection
config <- spark_config()
config$spark.executor.memory <- "8G"
config$spark.executor.cores <- 2
config$spark.executor.instances <- 3
config$spark.dynamicAllocation.enabled <- "false"

# initiate connection
sc <- spark_connect(master = "local", config = config, version = "2.4.3")

# unhash to verify version: spark_version(sc)

# select data for spark and create spark table
spark_data <- df %>% select(stars, user_id, business_id, name, city, categories)

yelp <- sdf_copy_to(sc, spark_data, "yelp", overwrite = TRUE)

# Transform features
yelp <- yelp %>% ft_string_indexer(input_col = "user_id", output_col = "user_index") %>% 
    ft_string_indexer(input_col = "business_id", output_col = "item_index") %>% 
    select(-user_id, -business_id) %>% sdf_register("yelp")

# randomly split / train test data
split <- sdf_random_split(yelp, training = 0.8, testing = 0.2, seed = 1)

# store training / test sets
train <- sdf_register(split$training, "train")
test <- sdf_register(split$testing, "test")

# tidy train for algoritms that require only user/item inputs
ui_train <- tbl(sc, "train") %>% select(user_index, item_index, stars)
ui_test <- tbl(sc, "test") %>% select(user_index, item_index, stars)

ALS

Once connected, we applied the alternating least squares (ALS) for our recommender predictions.

# build model using user/business/ratings
als_fit <- ml_als(ui_train, max_iter = 5, nonnegative = TRUE, rating_col = "stars", 
    user_col = "user_index", item_col = "item_index")

# predict from the model for the training data
als_predict_train <- ml_predict(als_fit, ui_train) %>% collect()
als_predict_test <- ml_predict(als_fit, ui_test) %>% collect()

# Remove NaN (result of test/train splits - not data)
als_predict_train <- als_predict_train[!is.na(als_predict_train$prediction), 
    ]
als_predict_test <- als_predict_test[!is.na(als_predict_test$prediction), ]

# View results
als_predict_test %>% head %>% kable() %>% kable_styling()

Metrics

Our ALS calculations for RMSE, MSE, and MAE can be viewed below:

# Calculate RMSE/MSE/MAE
als_mse_train <- mean((als_predict_train$stars - als_predict_train$prediction)^2)
als_rmse_train <- sqrt(als_mse_train)
als_mae_train <- mean(abs(als_predict_train$stars - als_predict_train$prediction))

als_mse_test <- mean((als_predict_test$stars - als_predict_test$prediction)^2)
als_rmse_test <- sqrt(als_mse_test)
als_mae_test <- mean(abs(als_predict_test$stars - als_predict_test$prediction))

# View metrics
cbind(als_rmse_train, als_mse_train, als_mae_train)

FALSE      als_rmse_train als_mse_train als_mae_train
FALSE [1,]      0.3614416       0.13064     0.2490331

cbind(als_rmse_test, als_mse_test, als_mae_test)

FALSE      als_rmse_test als_mse_test als_mae_test
FALSE [1,]      1.382284     1.910709     1.096935